JP7209905B1 - Water treatment system, aeration amount control device, and aeration amount control method - Google Patents

Abstract

The water treatment system (100) includes a water quality measurement value acquisition unit (50) that acquires a water quality measurement value at a first time point of the water to be treated, and water quality information previously acquired from time-series changes in water quality of the water to be treated. A water quality change pattern acquisition unit (40) for acquiring a plurality of water quality change patterns according to the conditions at the time of acquisition, and from the water quality change pattern of the water quality change pattern acquisition unit (40) to the acquisition condition of the water quality measurement value at the first time an inflow water quality estimating unit (30) for selecting a matching water quality variation pattern and estimating an inflow water quality value after the first time point based on the selected water quality variation pattern; and a control unit (60) for controlling the amount of aeration from the blower (20) to the reaction tank (10).

Description

TECHNICAL FIELD The present application relates to a water treatment system, an aeration amount control device, and an aeration amount control method.

There is an activated sludge method as a sewage treatment method containing organic matter and nitrogen. The activated sludge method is a method of oxidizing and decomposing contaminants in wastewater by accumulating microorganisms (activated sludge) with a purification function in a reaction tank and mixing and contacting this with wastewater and aerating it. . In order to sufficiently purify this contaminant, it is necessary to supply (aerate) an appropriate amount of air to the bioreactor. Moreover, since the residence time in the reaction tank is as long as about 10 hours, it is necessary to control the amount of aeration in response to fluctuations in the quality of the influent (the concentration of contaminants in the influent).

On the other hand, it is possible to control the amount of aeration using the quality of the influent, which is obtained by measuring it using a sensor that can continuously measure the quality of the inflow, or by using estimation from past time-series data. is known (see Patent Documents 1 and 2, for example).

Japanese Patent No. 6764487 JP-A-2005-125229

In Patent Literature 1, in order to continuously measure the quality of influent water, highly accurate measurement is achieved by arranging water quality sensors such as an ammonia meter, a total nitrogen meter, and a BOD meter. However, these measuring instruments are expensive and require frequent calibration and maintenance, so maintenance of the sensors is costly and labor intensive.

In Patent Literature 2, when determining a target value for water quality control, inflow water quality is predicted based on past time-series data. However, if the current measured inflow water quality has changed from the water quality at the time when the past data was acquired, there is a risk that the prediction accuracy will decrease.

The present application discloses a technique for solving the above problems, and estimates the quality of the inflow with high accuracy without using a large number of measuring instruments for continuously measuring the quality of the inflow, and detects fluctuations in the quality of the inflow. It is an object of the present invention to provide a water treatment system, an aeration amount control device, and an aeration amount control method capable of suppressing fluctuations in treated water quality by supplying a required amount of aeration without delay according to.

The water treatment system disclosed in the present application comprises:
In a water treatment system that aerates a reaction tank from a blower to perform water treatment by biological oxidation,
a first measured water quality acquisition unit that acquires a measured water quality value at a first point in time of the water to be treated that has flowed into the reaction tank;
a water quality variation pattern acquisition unit that acquires a plurality of water quality variation patterns according to the conditions at the time of acquiring the water quality information acquired in advance from time-series changes in the water quality information of the water to be treated;
Selecting one water quality change pattern that matches the conditions at the time of acquiring the water quality measurement value at the first time from the plurality of water quality change patterns provided by the water quality change pattern acquisition unit, and selecting the selected one water quality change pattern an inflow water quality estimation unit that estimates an inflow water quality value after the first time point based on the water quality measurement value at the first time point measured by the first water quality measurement value acquisition unit ;
and a control unit for controlling the aeration amount of the blower after the first point in time based on the inflow water quality value estimated by the inflow water quality estimation unit.

According to the water treatment system according to the present disclosure, the influent water quality value is estimated with high accuracy without using a measuring instrument that can continuously measure the influent water quality, and the necessary aeration amount is supplied according to the fluctuation of the influent water quality value. As a result, excessive aeration can be reduced while suppressing fluctuations in treated water quality.

1 is a block diagram showing the configuration of a water treatment system according to Embodiment 1; FIG. 1 is a block diagram showing an aeration amount control device according to Embodiment 1; FIG. FIG. 5 is a diagram for explaining an inflow water quality variation pattern according to Embodiment 1; FIG. 4 is a diagram showing an example of a water quality variation pattern selected by the aeration amount control device according to Embodiment 1; 4 is a diagram showing an operation flow of the aeration amount control device according to Embodiment 1; FIG. FIG. 4 is a block diagram showing an aeration amount control device according to Embodiment 2; FIG. 10 is a diagram showing an example of a water quality variation pattern selected by the aeration amount control device according to Embodiment 2; FIG. 11 is a block diagram showing the configuration of a water treatment system according to Embodiment 3; 3 is a hardware configuration diagram of an inflow water quality estimation unit and an aeration amount control device according to Embodiments 1 to 3. FIG. FIG. 10 is a diagram showing the configuration of an inflow water quality inference device according to Embodiment 4; FIG. 13 is a diagram showing the configuration of a learning device of an inflow water quality inference device according to Embodiment 4; 12 is a flow chart of learning using the learning device of FIG. 11; FIG. 12 is a diagram showing the configuration of an inference device according to Embodiment 4; FIG. 14 is a flowchart for inferring an inflow water quality value using the inference device of FIG. 13; FIG.

The present embodiment will be described below with reference to the drawings. In addition, in each figure, the same code|symbol shall show the same or a corresponding part.
Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the present specification and drawings, constituent elements having substantially the same functions are denoted by the same reference numerals, thereby omitting redundant description. In the drawings, figures showing the configuration of the device and the shapes of the members only show the schematic configuration and shape of the device and the members. The relative size and relative position of each member illustrated in each drawing do not necessarily accurately express the size relationship and positional relationship between actual members.

Embodiment 1.
A water treatment system according to Embodiment 1 will be described below with reference to FIGS. 1 to 5. FIG.
FIG. 1 is a diagram showing the configuration of a water treatment system according to Embodiment 1, and FIG. 2 is a diagram showing the configuration of an aeration amount control device. In FIG. 1, the water treatment system 100 includes an inflow portion 15 for introducing the water to be treated, an outflow portion 16 for discharging the treated water, a biological reaction tank 10 having diffuser plates 11, 12, 13 inside, and each diffuser. Air blower 20 for sending air to air plates 11, 12, 13, air volume control valves 71, 72, 73 for adjusting the amount of air from the blower 20, and aeration volume for controlling the air volume control valves 71, 72, 73 target aeration amount calculation units 61, 62, and 63 for calculating the target aeration amount calculation units 61, 62, and 63 for estimating the inflow water quality value to calculate the aeration amount; an inflow water quality measurement value acquisition unit 50 that acquires the measured value of the water quality of the water to be treated that has flowed into the inflow water quality estimation unit 30 and transmits it to the inflow water quality estimation unit 30; A variation pattern acquisition unit 40 is provided.

In addition, the air volume control valve 70 (71, 72, 73), the target aeration amount calculation unit 60 (61, 62, 63), the inflow water quality estimation unit 30, the inflow water quality measurement value acquisition unit 50, and the inflow water quality fluctuation pattern acquisition unit 40 The aeration amount control device is constructed. The air volume control valve and the target aeration volume calculation unit are denoted by the air volume control valve 70 and the target aeration volume calculation unit 60, respectively, when collectively indicated, and the air volume control valves 71 and 72 are used when individual parts are indicated. , 73, and target aeration amount calculators 61, 62, 63.
Details of each component will be described below.

The biological reaction tank 10 is a reaction tank in which activated sludge is stored. The air supplied from the blower 20 is supplied into the bioreactor 10 from the diffuser plates 11, 12, 13 through the pipe 20a.

The air volume control valves 71, 72 and 73 are installed respectively in the pipes branched from the pipe 20a to the diffuser plates 11, 12 and 13, respectively. By adjusting the opening degrees of the air volume control valves 71, 72, 73, the aeration volumes supplied to the diffuser plates 11, 12, 13 can be individually adjusted.

The target aeration amount calculators 61, 62, 63 calculate the target value of the aeration amount supplied from the air diffuser plates 11, 12, 13 for each arbitrary period, and calculate the target value of the aeration amount via the signal lines 61a, 62a, 63a. The value is sent to the air volume control valves 71 , 72 , 73 . The period for calculating the target value of the aeration rate is preferably between 1 second and 5 minutes, but can be set arbitrarily according to the characteristics of the plant. The air volume control valves 71 , 72 , 73 are opened so that the aeration volume supplied from the diffuser plates 11 , 12 , 13 becomes equal to the target value of the aeration volume calculated by the target aeration volume calculation units 61 , 62 , 63 . Adjust the intensity. In addition, although FIG. 1 shows an example in which three diffuser plates, three air volume control valves, and three target aeration volume calculation units are provided, the number of units is not limited to this. Depending on the scale of the biological reaction tank 10 and the characteristics of the plant, the number of diffuser plates, air volume control valves, and target aeration volume calculators can be arbitrarily changed.

The inflow part 15 is a pipe or water channel through which the water to be treated flows into the biological reaction tank 10 . The outflow part 16 is a pipe or water channel through which the treated water treated in the biological reaction tank 10 flows out of the biological reaction tank 10 . The water to be treated that has flowed in from the inflow part 15 comes into contact with air and activated sludge supplied by aeration in the biological reaction tank 10, thereby promoting oxidation and decomposition of contaminants in the water to be treated, and is discharged as treated water. will be

Next, the inflow water quality variation pattern acquisition unit 40 will be described.
The influent water quality fluctuation pattern acquisition unit 40 receives and holds a time-variation pattern of contaminant concentration obtained in advance by arranging the contaminant concentration in the water to be treated in chronological order according to the acquisition time. One or more types of contaminants are selected from the objects to be treated in the biological reaction tank 10 . Examples thereof include BOD (Biochemical Oxygen Demand), COD (Chemical Oxygen Demand), ammonium nitrogen, total nitrogen, Kjeldahl nitrogen, total phosphorus, and phosphate phosphorus. In addition, it is possible to input a plurality of inflow water quality variation patterns according to the type of contaminants or the characteristics of the date and time.

FIG. 3 is a diagram for explaining the inflow water quality variation pattern acquired by the inflow water quality variation pattern acquisition unit 40. As shown in FIG. A typical sewage treatment plant has a certain pattern of variation in influent quality over the course of a day. The fluctuation pattern differs from station to station, and there is a fluctuation pattern in which peaks are seen in the morning and evening, which is characteristic of residential areas, and a fluctuation pattern in which peaks are seen only in the daytime, which is characteristic of office areas. Sampling the influent water for a day at intervals of 1 or 2 hours and measuring the water quality. normalize the influent water quality values at other times by the values.

In FIG. 3 , the vertical axis represents the concentration of contaminants in any water to be treated, and is an example of the change over time when normalized with the inflow water quality value at 9:00 am on a certain day as 1, that is, the variation pattern of water quality. As described above, the variation pattern of influent water quality depends on the activity pattern of people in the treatment area, so the variation pattern of influent water quality may be different on weekdays and weekends, for example. For this reason, the variation patterns of influent water quality are obtained separately for weekdays and weekends in advance. It is input to the variation pattern acquisition unit 40 .

It should be noted that dividing the inflow water quality fluctuation pattern by weekdays and weekends is just one example, and according to the characteristics of the treatment area, multiple indices such as month, day of the week, and the operation rate of the factory in the treatment area can be used to determine the inflow water quality fluctuation pattern. You can divide. In addition, the inflow water quality fluctuation pattern does not necessarily have to be a day unit, and a pattern period such as 1 minute, 1 hour, 1 week, 1 month, etc. may be selected according to the fluctuation cycle of the inflow water quality at each station. . The inflow water quality variation pattern is input to the inflow water quality variation pattern acquisition unit 40 for each period of the set pattern. That is, when the inflow water quality variation pattern is for one day, the inflow water quality variation pattern is input to the inflow water quality variation pattern acquisition unit 40 every day.

The inflow water quality measurement value acquiring unit 50 stores the instantaneous value of the contaminant concentration measured by sampling the inflow water spotwise in the inflow unit 15 or in an area close to the inflow unit 15 in the biological reaction tank 10, together with the sampling date and time. entered together. The type of contaminants input to the influent water quality measurement value acquisition unit 50 is preferably the same as the type of contaminants input to the inflow water quality variation pattern acquisition unit 40 . The method of inputting the inflow water quality measurement value to the inflow water quality measurement value acquisition unit 50 may be any method such as tablet, mouse operation, keyboard operation, or screen input of the central monitoring system.

The frequency of inputting the inflow water quality measurement value to the inflow water quality measurement value acquisition unit 50 is arbitrary. It is desirable to input the inflow water quality measurement value to the inflow water quality measurement value acquisition unit 50 at a frequency equal to or less than the period of . For example, when the influent water quality fluctuation pattern input to the influent water quality fluctuation pattern acquisition unit 40 is in units of one day (24 hours), the influent water quality measurement value acquisition unit 50 receives measured values at least once a day. Input is preferred.

The inflow water quality estimation unit 30 estimates the inflow water quality variation pattern transmitted from the inflow water quality variation pattern acquisition unit 40 via the signal line 40a and the inflow water quality transmitted from the inflow water quality measurement value acquisition unit 50 via the signal line 50a. Based on the measured value, the influent water quality value after the sampling date and time of the influent water quality measurement is estimated. The inflow water quality value estimated by the inflow water quality estimation unit 30 is transmitted to the target aeration amount calculation unit 60 via the signal line 30a. The target aeration amount calculator 60 calculates the target value of the aeration amount based on the influent water quality value estimated by the influent water quality estimator 30 .

Next, the operation of the inflow water quality estimation unit 30 will be described. The inflow water quality estimation unit 30 includes a sampling date/time extraction unit 31 , an inflow water quality variation pattern selection unit 32 , and an inflow water quality calculation unit 33 . The sampling date and time extraction unit 31 extracts the characteristics of the sampling date and time of the influent water quality measurement value input to the inflow water quality measurement value acquisition unit 50, and extracts the date and time of the inflow water quality change pattern held in the inflow water quality change pattern acquisition unit 40. Associate with features. For example, when inflow water quality fluctuation patterns corresponding to weekdays and holidays are respectively input to the inflow water quality fluctuation pattern acquisition unit 40, and the inflow water quality measurement values of weekdays in the inflow water quality measurement value acquisition unit 50 are input, The sampling date and time extraction unit 31 extracts "weekday" as a feature of the sampling date and time.

The inflow water quality variation pattern selection unit 32 selects an inflow water quality variation pattern corresponding to the characteristics of the sampling date and time extracted by the sampling date and time extraction unit 31 from the inflow water quality variation patterns input to the inflow water quality variation pattern acquisition unit 40. For example, in the above example, when "weekdays" are extracted as a feature of the sampling date and time, the influent water quality variation pattern corresponding to "weekdays" is selected.

The inflow water quality calculation unit 33 determines the sampling date and time for measuring the inflow water quality based on the inflow water quality variation pattern transmitted from the inflow water quality variation pattern selection unit 32 and the inflow water quality measurement value input to the inflow water quality measurement value acquisition unit 50. Calculate subsequent influent water quality values. That is, if the sampling date and time of the inflow water quality measurement is set to the first time, the inflow water quality value after the first time is calculated.

The estimated value of the inflow water quality calculated by the inflow water quality calculator 33 will be described with reference to FIG. For example, the fluctuation pattern of the inflow total nitrogen concentration on a daily basis is input to the inflow water quality fluctuation pattern acquisition unit 40, and 25 mg/L is input to the inflow water quality measurement value acquisition unit 50 as the inflow total nitrogen concentration at 9:00 on October 1. In this case, the inflow water quality calculation unit 33 calculates the value of the inflow total nitrogen concentration at each time from the measured value of 25 mg / L at 9:00 on October 1st to 9:00 on October 2nd one day later. Output. That is, in FIG. 4, the variation pattern indicated by the solid line is calculated.

For example, if the sampling period is set to 8 hours, and the inflow water quality measurement value is input from the inflow water quality measurement value acquisition unit 50 at 17:00 on October 1, the same weekday, the inflow water quality variation pattern selection unit 32 selects the pattern is not changed, but the value of the inflow total nitrogen concentration at each time until 9:00 on October 2, one day later, is re-estimated according to the input inflow water quality measurement value. Similarly, re-estimation will also be performed for the inflow water quality measurement value input at 1:00 on October 2, eight hours later.

Here, the estimation algorithm installed in the inflow water quality calculation unit 33 may be any method as long as it adopts a method that can estimate the time fluctuation of the inflow water quality using the inflow water quality fluctuation pattern and the inflow water quality measurement value as inputs. Examples include linear/nonlinear regression models, machine learning, reinforcement learning, deep reinforcement learning, deep learning, random forests, neural networks, and other prediction methods using artificial intelligence. These inflow-water quality variation patterns to be estimated can be input to the inflow-water-quality variation pattern acquisition unit 40 as learned inflow-water quality variation patterns.

If a variation pattern of the inflow total nitrogen concentration on a weekly basis is input to the inflow water quality variation pattern acquisition unit 40, the inflow total nitrogen concentration at each time from 9:00 on October 1 to 9:00 on October 8 is obtained. Output.

As described above, since the variation pattern of the inflow water quality is obtained in advance, there is no need to use a sensor capable of continuously measuring the inflow water quality. Furthermore, by estimating the inflow water quality after the sampling date and time of the inflow water quality measurement in combination with the inflow water quality measurement value, it is possible to improve the accuracy of the inflow water quality estimation for each sampling.

The target aeration amount calculation unit 60 ( 61 , 62 , 63 ) calculates the target value of the aeration amount supplied from the diffuser plates 11 , 12 , 13 based on the inflow water quality value estimated by the inflow water quality estimation unit 30 . The target aeration amount calculation units 61, 62, and 63 calculate the target aeration amount in conjunction with the inflow water quality estimated value corresponding to the time when the target aeration amount is calculated. For example, it is calculated based on the following equations (1) to (3).
Qair1=K1×TN (1)
Qair2=K2×TN (2)
Qair3=K3×TN (3)
Here, Qair1, Qair2, and Qair3 are target aeration amounts calculated by the target aeration amount calculation units 61, 62, and 63, TN is the inflow water quality estimated by the inflow water quality estimation unit 30, and K1, K2, and K3 are proportional constants. be.

By calculating the target aeration rate based on the formulas (1) to (3), it becomes possible to control the aeration rate following fluctuations in the inflow water quality. As a result, by suppressing the control delay, fluctuations in the water quality after treatment are also suppressed, and while good treated water quality is obtained, the excessive aeration amount can be reduced. It should be noted that K1, K2, and K3 do not all have to be the same value, and arbitrary values can be set according to the positions of the diffuser plates 11, 12, and 13. FIG. In particular, when it is desired to minimize the influence of fluctuations in the quality of influent water, the fluctuations in the quality of influent water can be quickly suppressed on the upstream side of the biological reaction tank 10 by setting the respective proportionality constants so that K1>K2>K3. be able to. Furthermore, equations (1) to (3) are examples, and if it is desired to control the amount of aeration following the inflow load (inflow contaminant concentration x inflow water amount), an inflow water amount term may be added.

Next, the operation flow of the aeration amount control device according to Embodiment 1 will be described with reference to FIG.
Inflow water quality variation patterns are input to the inflow water quality variation pattern acquisition unit 40 in advance. From this state, the control of the aeration amount control device is started.
The operation flow of the aeration amount control device includes six steps, ST1 to ST6, shown in the figure. When the aeration amount control is started, the flow of steps ST1 to ST6 is repeated at certain time intervals Δt. Here, Δt is set at an interval equal to or less than the update cycle T of the target aeration amount of the target aeration amount calculator 60, and Δt is preferably in the range of about 1 second to 1 minute. However, it is necessary to set a time Δt longer than the time required to process all steps ST1 to ST6.

First, in step ST1, it is determined whether or not the time t is the update period of the target aeration amount of the target aeration amount calculator 60 . When the update cycle T of the target aeration amount is 5 minutes, step ST1 is determined to be Yes at 5-minute intervals. If Yes in step ST1, the process proceeds to the next step ST2, and if No in step ST1, the process proceeds to step ST5.

In step ST2, it is determined whether or not a new measurement value of the inflow water quality is input to the inflow water quality measurement value acquiring unit 50 between t-Δt, which is one cycle before the repetition interval Δt, and time t. When the input cycle of the inflow water quality to the inflow water quality measurement value acquisition unit 50 is one day, step ST2 is determined to be Yes at intervals of one day. If step ST2 is Yes, the process proceeds to the next step ST3, and if step ST2 is No, the process proceeds to step ST5.

Here, the water quality is measured after sampling water for measuring the quality of inflow water, and the result is input to the measured value acquisition unit 50 for quality of inflow water. Note that it is more past than t. For example, sampling for inflow water quality measurement is performed at 9:00 on October 1, and after one hour for water quality analysis, the measured value at 9:00 on October 1 is input to the inflow water quality measurement value acquisition unit 50. In this case, when the time t in the operation flow of the aeration amount control device is 10:00 on October 1st, the inflow water quality measurement value at 9:00 on October 1st is input.

Next, in step ST3, the inflow water quality estimating unit 30 estimates the inflow water quality value after the sampling date and time of the inflow water quality measurement. In this embodiment, as shown in FIG. 4, the inflow water quality fluctuation pattern is input to the inflow water quality fluctuation pattern acquisition unit 40 on a daily basis. The variation of influent water quality values over time is estimated.

Next, in step ST4, the inflow water quality estimated value recorded at time t-Δt is updated to the inflow water quality estimated value estimated in step ST3.

In step ST5, the target aeration amount calculators 61, 62, and 63 refer to the inflow water quality estimated value at time t. In the present embodiment, when the time t is 10:00 on October 1st, the inflow water quality estimated value at 10:00 on October 1st is referred to.

In step ST6, the target aeration amount at time t in each of the target aeration amount calculators 61, 62, and 63 is calculated using the inflow water quality estimated value and formulas (1) to (3) referred to in step ST5. The calculated target aeration amount is transmitted to the air volume control valves 71, 72, 73 via signal lines 61a, 62a, 63a, and each air volume control valve 71, 72, 73 adjusts the valve so as to achieve the target aeration volume. Adjust the opening.

In the operation flow of the aeration amount control device of FIG. 5, the water quality of the water to be treated that has flowed into the biological reaction tank 10 is sampled at 9:00 on October 1, which is the first time, and the water quality is measured. After that, at the second time, 10:00 on October 1, the inflow water quality measurement value acquisition unit 50 acquires the inflow water quality measurement value at the first time, and the inflow water quality estimation unit 30 calculates the inflow water quality after the sampling date and time of the inflow water quality measurement. to estimate Here, since the water quality measurement value acquisition cycle for the inflow water quality measurement value acquisition unit 50 is one day, the water quality of the water to be treated flowing into the biological reaction tank 10 is sampled at 9:00 on October 2, which is the third time. and measure the water quality. After that, at the fourth time, 10:00 on October 2, the influent water quality measurement value acquisition unit 50 acquires the inflow water quality measurement value at the third time, and the inflow water quality estimation unit 30 outputs the sampling date and time of the inflow water quality measurement (the third time ) to estimate the influent water quality after In this way, sampling of inflow water quality, acquisition of measured values, and estimation of inflow water quality are repeated. Note that the third time may be before 9:00 on October 2 if sampling and measurement values are determined so that inflow water quality measurement values after the first time can be obtained at the fourth time. . Of course, the closer to 9:00 on October 2nd, the more recent influent water quality readings will be available.

As described above, according to the water treatment system according to Embodiment 1, the influent water quality variation pattern input to the influent water quality variation pattern acquisition unit 40 and the influent water quality measurement value input to the inflow water quality measurement value acquisition unit 50 By estimating the inflow water quality value in the inflow water quality estimating unit 30 based on, it is possible to estimate the highly accurate inflow water quality value that reflects both the difference in the inflow water quality fluctuation pattern depending on the sampling date and the most recent water quality measurement value. . Furthermore, using the estimated inflow water quality value, the target aeration amount is calculated in the target aeration amount calculation units 61, 62, 63, and the air volume control valves 71, 72, 73 are adjusted according to the calculated target aeration amount. Therefore, it is possible to suppress the control delay and perform treatment according to the water quality value of the influent, and it is possible to reduce excessive aeration while obtaining good treated water quality.

Conventionally, the operator who measures the water quality and the operator who controls the operation of the water treatment system were separate, and it took time to reflect the water quality result in the control of the water treatment system, but this embodiment By controlling the water treatment system using the estimated inflow water quality estimated value as described above, it is possible to control according to the water quality.

Embodiment 2.
A water treatment system according to Embodiment 2 will be described below with reference to FIG.
FIG. 6 is a block diagram showing the configuration of an aeration amount control device for a water treatment system according to Embodiment 2. As shown in FIG. The difference from the first embodiment is that the inflow water quality estimation unit 30 is provided with a rainfall effect determination unit 34 . Since other configurations are the same as those of the first embodiment, description thereof is omitted.

In FIG. 6, the rainfall effect determination unit 34 determines whether or not the inflow water quality measurement value input to the inflow water quality measurement value acquisition unit 50 is affected by rain. The water quality measurement value and the rainfall influence degree are transmitted to the inflow water quality calculation unit 33 .

When it rains, the water to be treated is diluted with rainwater, so the concentration of contaminants in the inflow becomes smaller than when it is sunny, and there is a concern that the accuracy of estimating the quality of the inflow deteriorates. Therefore, when the inflow water quality measurement value is lower than the determination threshold preset inside the rainfall effect determination unit 34, the degree of rainfall effect is calculated. For calculation of the degree of influence of rainfall, weather information such as rain cloud radar in the target processing area of the airport may be obtained and the result of judging the influence of rain may be used.

The degree of influence of rainfall is the ratio of dilution of inflow water by rainwater intrusion, and is calculated by Equation (4).
Dilution ratio = 1-Ssp_r/Ssp_ave (4)
Here, Ssp_r is an inflow water quality measurement value sampled on a date and time affected by rainfall, and Ssp_ave is an average value of the inflow water quality measurement values in fine weather. As for the average value of the inflow water quality measurement values, the inflow water quality measurement values are acquired a plurality of times on the same sampling date and time, and the average value is input to the rainfall effect determination unit 34 .

Next, the estimated value of the inflow water quality calculated by the inflow water quality calculation unit 33 when the rainfall effect determination unit 34 determines that there is rainfall will be described with reference to FIG.
For example, the fluctuation pattern of the inflow total nitrogen concentration on a daily basis is input to the inflow water quality fluctuation pattern acquisition unit 40, and 5 mg/L is input to the inflow water quality measurement value acquisition unit 50 as the inflow total nitrogen concentration at 9 o'clock on October 8. and the rainfall effect determination unit 34 determines that there is rainfall effect. In this case, in the inflow water quality calculation unit 33, the value of the inflow total nitrogen concentration at each time from the inflow water quality measurement value of 5 mg/L at 9:00 on October 8 to 9:00 on October 9 one day later. to output Here, the estimated value of the inflow total nitrogen concentration is smaller than that in the case where there is no influence of rainfall, and the fluctuation range of the inflow water quality value is also estimated to be smaller according to the dilution ratio by rainwater.

The dashed line in FIG. 7 is the variation pattern of the estimated inflow total nitrogen concentration when it is sunny on weekdays, that is, when there is no rainfall effect. is estimated and the pattern shown by the solid line can be obtained. In this way, the inflow water quality value can be accurately estimated even on a day affected by rainfall. The estimation algorithm installed in the inflow water quality calculation unit 33 may adopt a method that can estimate the time fluctuation of the inflow water quality value with the input of the inflow water quality fluctuation pattern, the inflow water quality measurement value, and the dilution ratio due to rainfall. The same methods as in Embodiment 1, such as nonlinear regression models, machine learning, reinforcement learning, deep reinforcement learning, deep learning, random forests, neural networks, and other prediction methods using artificial intelligence, can be used.

The target aeration amount calculators 61 , 62 and 63 calculate the target values of the aeration amounts to be supplied from the diffuser plates 11 , 12 and 13 based on the influent water quality values estimated by the influent water quality estimator 30 . The target aeration amount calculation units 61, 62, and 63 calculate the target aeration amount in conjunction with the inflow water quality estimated value corresponding to the time when the target aeration amount is calculated. In general, when there is rainfall, the concentration of contaminants in the influent is very low, so the same amount of aeration as in fine weather results in excessive aeration. In the present embodiment, the target aeration amount calculation units 61, 62, and 63 calculate the target aeration amount based on the estimated value of the inflow water quality that takes into account the influence of rainfall. In addition to stabilizing the air temperature, excessive aeration can be further suppressed compared to when the weather is fine.

As described above, according to the second embodiment, the inflow water quality is estimated using the degree of influence of rainfall determined by the rainfall effect determination unit 34 in the inflow water quality estimation unit 30. Even when an inflow water quality variation pattern different from that in fine weather is predicted, it is possible to estimate the inflow water quality value with high accuracy. Furthermore, target aeration amounts are calculated in target aeration amount calculation units 61, 62, and 63 using the estimated value of inflow water quality that takes into account the influence of rainfall, and the air volume control valves 71 and 72 are calculated according to the calculated target aeration amounts. , 73 can be adjusted according to the quality of the inflow water even when it rains, and it is possible to reduce excessive aeration while obtaining good treated water quality.

Embodiment 3.
A water treatment system according to Embodiment 3 will be described below with reference to FIG.
FIG. 8 is a block diagram showing the configuration of a water treatment system according to Embodiment 3. As shown in FIG. The difference from the first embodiment is that the contaminant concentration measuring unit 80 for measuring the concentration of contaminants at the time when the water to be treated that has flowed into the biological reaction tank 10 is finished being treated in the biological reaction tank 10 It is to have a measuring instrument. Since other configurations are the same as those in the first or second embodiment, description thereof is omitted.

The contaminant concentration measurement unit 80 according to Embodiment 3 corresponds to a second water quality measurement value acquisition unit different from the influent water quality measurement value acquisition unit 50 . Since the main role of the contaminant concentration measuring unit 80 is to measure the contaminant concentration at the time when the treatment is completed, the concentration measuring device should be installed at a position closer to the outflow part 16 in the biological reaction tank 10. is desirable. Alternatively, the concentration measuring device of the contaminant concentration measuring section 80 may be installed in the outflow section 16 .

One or more kinds of concentration measuring devices are selected from the objects to be treated in the biological reaction tank 10 . For example, there are BOD, COD, ammonium nitrogen, total nitrogen, Kjeldahl nitrogen, total phosphorus, phosphorous phosphorus, and the like. The contaminant concentration measuring unit 80 does not necessarily need to have a concentration measuring device of the same type as the water quality input to the influent water quality fluctuation pattern acquiring unit 40 and the influent water quality measurement value acquiring unit 50, and the concentration measuring device to be monitored as the treated water quality is selected. good to do
Contaminant concentrations measured by the contaminant concentration measurement unit 80 are transmitted to the target aeration amount calculation units 61, 62, 63 via the signal line 80a. Target aeration amount calculators 61, 62, and 63 calculate target values of aeration amounts based on the estimated value of the influent water quality estimated by the influent water quality estimation unit 30 and the contaminant concentration measured by the contaminant concentration measurement unit 80. calculate.

Next, the calculation method of the target value of the aeration amount in the target aeration amount calculators 61, 62 and 63 will be described. Hereinafter, the fluctuation pattern of the influent total nitrogen concentration and the influent water quality measurement value are input to the influent water quality fluctuation pattern acquisition unit 40 and the influent water quality measurement value acquisition unit 50, respectively, and the contaminant concentration measurement unit 80 uses the ammonia concentration meter as a contaminant concentration meter. A case in which a nitrogen concentration meter is provided will be described as an example.

The target value of the aeration amount is controlled to follow the fluctuation of the influent water quality value estimated by the influent water quality estimation unit 30, and operates so that the contaminant concentration measured by the contaminant concentration measurement unit 80 becomes equal to the target water quality. It is determined based on control (PI control: proportional integral control). As a specific example, the target value of the aeration amount is determined based on the following formulas (5) to (7). Of the components of the control unit in the third embodiment, the target water quality setting unit for setting the target water quality is not illustrated, but the target aeration amount calculation units 61, 62, and 63 have target water quality setting units. It is built in and the operator can set the target water quality from the outside.
Qair1=K1×TN+Kp1×(NH4−NH4*)
+Ki1×Σ(NH4−NH4*) (5)
Qair2=K2×TN+Kp2×(NH4−NH4*)
+Ki2×Σ(NH4−NH4*) (6)
Qair3=K3×TN+Kp3×(NH4−NH4*)
+Ki3×Σ(NH4−NH4*) (7)

Here, Qair1, Qair2, and Qair3 are the target aeration amounts calculated by the target aeration amount calculation units 61, 62, and 63, TN is the estimated value of the influent total nitrogen concentration estimated by the influent water quality estimation unit 30, K1, K2 and K3 are proportional constants, Kp1, Kp2 and Kp3 are proportional gains (constants), Ki1, Ki2 and Ki3 are integral gains (constants), NH4 is the ammonia nitrogen concentration measured by the contaminant concentration measurement unit 80, and NH4*. is the target value of ammonia nitrogen concentration. Also, Σ means the sum of the measured values of (NH4-NH4*) after the start of calculation of the aeration amount by the formulas (5) to (7). For example, considering the case where the calculation of the aeration amount based on formulas (5) to (7) is performed at intervals of one minute, the value of Σ(NH4−NH4*) after one hour is expressed by formulas (5) to It is the total value of 60 measurements of (NH4-NH4*) every minute from immediately after starting the calculation of the aeration amount by the formula (7).

The first term of the equations (5) to (7) is proportional control for the estimated value of the influent total nitrogen concentration estimated by the influent water quality estimating unit 30, whereby the aeration amount is adjusted according to the fluctuation of the influent water quality. control becomes possible.

The second and third terms of equations (5) to (7) are PI control based on the difference between the ammonia nitrogen concentration measured by the contaminant concentration measurement unit 80 and the target value of ammonia nitrogen, As a result, the amount of aeration is controlled so that the quality of the treated water is constant with respect to the target value, so that the necessary amount of aeration can be supplied to the biological reaction tank 10 in just the right amount. The proportional constants (K1, K2, K3), proportional gains (Kp1, Kp2, Kp3), and integral gains (Ki1, Ki2, Ki3) for the estimated value of the inflow water quality are all The values do not have to be equal, and arbitrary values can be set according to the positions of the diffuser plates 11 , 12 , and 13 . In particular, when it is desired to suppress the influence of fluctuations in the influent water quality value as much as possible, by setting the respective proportionality constants so that K1 > K2 > K3, the fluctuations in the influent water quality value can be made early on the upstream side of the biological reaction tank 10. can be suppressed.

Also, if you want to suppress the fluctuation of the treated water quality value as much as possible, by setting Kp1<Kp2<Kp3 and Ki1<Ki2<Ki3, the closer the air volume control valve is to the place where the contaminant concentration measurement unit 80 is installed, the more PI control is performed. can strengthen the influence of Furthermore, equations (5) to (7) are examples, and if you want to control the amount of aeration following the inflow load (concentration of inflow pollutants x amount of inflow water), you may add a term for the amount of inflow water to the first term. .

As described above, according to Embodiment 3, the target aeration amount calculation unit 61, By calculating the target aeration amount at 62 and 63, the responsiveness of the aeration amount to fluctuations in the inflow water quality can be improved without using a sensor capable of continuously measuring the inflow water quality, and the treated water quality can be controlled to be constant. . As a result, it is possible to suppress the fluctuation of the treated water quality value and obtain good treated water quality, and to reduce the excessive amount of aeration by supplying just enough aeration.

In the first to third embodiments described above, the influent water quality estimation unit 30 and the target aeration amount calculation unit 60 of the water treatment system 100 are configured from the processor 301 and the storage device 302 as shown in FIG. Configured. Although not shown, the storage device includes a volatile storage device such as a random access memory and a non-volatile auxiliary storage device such as a flash memory. Also, an auxiliary storage device such as a hard disk may be provided instead of the flash memory. Processor 301 executes a program input from storage device 302 . In this case, the program is input from the auxiliary storage device to the processor 301 via the volatile storage device. Further, the processor 301 may output data such as calculation results to the volatile storage device of the storage device 302, or may store the data in an auxiliary storage device via the volatile storage device.

Embodiment 4.
A water treatment system according to Embodiment 4 will be described below with reference to FIGS. 10 to 14. FIG. In this embodiment, an inflow water quality inference device configured by mounting a machine learning device on the inflow water quality estimation unit 30 shown in Embodiments 1 to 3 will be described.
FIG. 10 is a diagram showing the configuration of an inflow water quality inference device 400 of a water treatment system according to Embodiment 4, which includes a learning device 410 and an inference device 420 . The inflow water quality inference device 400 may be the inflow water quality estimation unit 30 . The procedure for inferring the inflow water quality value will be described below by dividing it into a "learning phase" and an "utilization phase" in which the inference is actually made.

<Learning phase>
FIG. 11 is a diagram showing the configuration of the learning device 410. As shown in FIG. The learning device 410 includes a data acquisition unit 411 , a model generation unit 412 and a learned model storage unit 413 .
FIG. 12 is a flow chart showing a processing procedure for executing the learning phase using the learning device 410. As shown in FIG.

The data acquisition unit 411 acquires the pollutant type and concentration data b1 and the measurement date and time data b2 from the influent water quality measurement value acquisition unit 50 as input data, and combines them to obtain time-series learning data (step ST101).

The model generator 412 learns fluctuations in the inflow water quality based on the learning data output from the data acquisition unit 411 (step ST102). That is, from the time-series data of the concentration of a plurality of pollutants for a certain period, for example, water quality fluctuations by day (weekdays or holidays) and by type of polluted part are learned, and the learned model 414, which is a water quality fluctuation pattern, is learned. to generate For learning to generate the learned model 414, for example, if linear/nonlinear regression models, machine learning, reinforcement learning, deep reinforcement learning, deep learning, random forest, neural networks, and other prediction methods using artificial intelligence are used. good.

The model generation unit 412 generates and outputs the learned model 414 by executing the learning as described above, and the learned model storage unit 413 stores the learned model 414 output from the model generation unit 412. (Step ST103).

<Utilization phase>
FIG. 13 is a diagram showing the configuration of the inference device 420. As shown in FIG. The inference device 420 includes a data acquisition unit 421 and an inference unit 422 .
FIG. 14 is a flow chart showing a processing procedure for executing the utilization phase, in which the inference device 420 is used to infer the water quality value of the influent.

The data acquisition unit 421 acquires the pollutant water quality measurement data b11 from the inflow water quality measurement value acquisition unit 50 and the inflow water quality change pattern b12 selected by the inflow water quality change pattern selection unit 32 as input data (step ST111).

The inference unit 422 uses the learned model 414 to infer changes in the inflow water quality. That is, by inputting the new pollutant water quality measurement data b11 acquired by the data acquisition unit and the selected influent water quality fluctuation pattern b12 at a certain point into this learned model 414 (step ST112), the water quality fluctuation value can be inferred with high accuracy and output from inference device 420 (step ST113).
The result of the inferred inflow water quality value is output to the target aeration amount calculation unit 60 (step ST114).

The inflow water quality measurement data b11 from the inflow water quality measurement value acquisition unit 50 input from the data acquisition unit 421 and the inflow water quality change pattern b12 selected by the inflow water quality change pattern selection unit 32 are input to the inference unit 422. However, the influent water quality variation pattern b12 selected by the influent water quality variation pattern selection unit 32 may or may not be used as reference data. The water quality measurement data b11 of pollutants acquired from the inflow water quality measurement value acquisition unit 50 is attached with water quality information such as the date and time of measurement and the type of pollutants, and can be applied to the learned data. Because it is possible.

In the above description, an example corresponding to the first embodiment has been described. It goes without saying that 400 can be built.
Further, it goes without saying that the third embodiment can also be applied.

The reasoning device 420 may be the inflow water quality estimation unit 30, and the learning device 410 and the learned model storage unit 413 may be externally attached to the inflow water quality estimation unit 30 of Embodiments 1 to 3, or as shown in FIG. may be integrated as the influent water quality inference device 400 in the .

An example of the hardware of the influent water quality inference device 400 is composed of a processor and a storage device as in the configuration shown in FIG. The processor realizes each function of the learning device 410 and the inference device 420 of the influent water quality inference device 400 by executing the program.

The pattern input to the inflow water quality variation pattern acquisition unit 40 of Embodiments 1 to 4 may be an inflow water quality variation pattern using a learned model.

In the fourth embodiment, the learned model 414 learned by the model generation unit 412 is used to output the inference result of the inflow water quality. You may make it output the inference result of an inflow water quality based on.
Moreover, although the learning phase and the utilization phase are described separately in the fourth embodiment, the learning phase may be performed first, and then the utilization phase may be performed, or both may be performed in parallel. When both are performed in parallel, some threshold value for completion of learning such as the number of data acquisitions is set, and only learning is performed while learning is incomplete, and both are performed in parallel after completion of learning.

As described above, according to the fourth embodiment, in addition to the effects of the first embodiment, the type and concentration data b1 of contaminants and the measurement date and time data b2 are acquired from the influent water quality measurement value acquisition unit 50, and both are combined to generate a learned model as time-series learning data, and by inputting the new current pollutant water quality measurement data b11 and the selected inflow water quality fluctuation pattern b12 at a certain point in this learned model, Since it is possible to infer the water quality fluctuation value with high accuracy, it is possible to control the amount of aeration with higher accuracy based on the result.

While this disclosure describes various exemplary embodiments and examples, various features, aspects, and functions described in one or more of the embodiments may vary from particular embodiment to embodiment. The embodiments are applicable singly or in various combinations without being limited to the application.
Accordingly, numerous variations not illustrated are envisioned within the scope of the technology disclosed herein. For example, modification, addition or omission of at least one component, extraction of at least one component, and combination with components of other embodiments shall be included.

10: biological reaction tank 11, 12, 13: diffuser plate 15: inflow part 16: outflow part 20: blower 20a: pipe 30: inflow water quality estimation part 30a: signal line 31: sampling date and time Extraction unit 32: Inflow water quality fluctuation pattern selection unit 33: Inflow water quality calculation unit 34: Rainfall effect determination unit 40: Inflow water quality fluctuation pattern acquisition unit 40a: Signal line 50: Inflow water quality measurement value acquisition unit 50a : signal line 60, 61, 62, 63: target aeration amount calculation unit 61a, 62a, 63a: signal line 70, 71, 72, 73: air volume control valve 80: contaminant concentration measurement unit 80a: signal Line 100: Water treatment system 301: Processor 302: Storage device 400: Influent water quality inference device 410: Learning device 411: Data acquisition unit 412: Model generation unit 413: Trained model storage unit 414 : learned model, 420: inference device, 421: data acquisition unit, 422: inference unit.

Claims (8)

In a water treatment system that aerates a reaction tank from a blower to perform water treatment by biological oxidation,
a first measured water quality acquisition unit that acquires a measured water quality value at a first point in time of the water to be treated that has flowed into the reaction tank;
a water quality variation pattern acquisition unit that acquires a plurality of water quality variation patterns according to the conditions at the time of acquiring the water quality information acquired in advance from time-series changes in the water quality information of the water to be treated;
Selecting one water quality change pattern that matches the conditions at the time of acquiring the water quality measurement value at the first time from the plurality of water quality change patterns provided by the water quality change pattern acquisition unit, and selecting the selected one water quality change pattern an inflow water quality estimation unit that estimates an inflow water quality value after the first time point based on the water quality measurement value at the first time point acquired by the first water quality measurement value acquisition unit ;
a control unit that controls the aeration amount of the blower after the first time based on the inflow water quality value estimated by the inflow water quality estimation unit;
water treatment system with a second water quality measurement value acquisition unit, wherein the second water quality measurement value acquisition unit acquires the water quality measurement value of the treated water treated in the reaction tank;
2. The control unit according to claim 1, wherein the control unit further controls the aeration amount of the blower after the first time based on the water quality measurement value of the treated water acquired by the second water quality measurement value acquisition unit. water treatment system. The inflow water quality estimation unit
A data acquisition unit that acquires in advance a time-series change in the water quality information of the water to be treated as learning data; A learning device comprising a model generation unit that generates a trained model for inferring a water quality variation pattern; 3. A water treatment system according to claim 1 or 2, comprising an inference device for inferring influent water quality values after said first time point from water quality measurements at one time point. An aeration amount control device used in a water treatment system that aerates a reaction tank from a blower to perform water treatment by biological oxidation,
a first measured water quality acquisition unit that acquires a measured water quality value at a first point in time of the water to be treated that has flowed into the reaction tank;
a water quality variation pattern acquisition unit that acquires a plurality of water quality variation patterns according to the conditions at the time of acquiring the water quality information acquired in advance from time-series changes in the water quality information of the water to be treated;
Selecting one water quality change pattern that matches the conditions at the time of acquiring the water quality measurement value at the first time from the plurality of water quality change patterns provided by the water quality change pattern acquisition unit, and selecting the selected one water quality change pattern an inflow water quality estimation unit that estimates an inflow water quality value after the first time point based on the water quality measurement value at the first time point acquired by the first water quality measurement value acquisition unit ;
a target aeration amount calculation unit that calculates a target aeration amount to be supplied from the blower to the reaction tank after the first time based on the influent water quality value estimated by the inflow water quality estimation unit. . The first water quality measurement value acquisition unit acquires the water quality measurement value of the water to be treated that has flowed into the reaction tank at a second time point after the first time point,
The inflow water quality estimation unit
Select the water quality change pattern from the water quality change pattern acquisition unit corresponding to the acquisition date and time of the water quality measurement value at the second time point acquired by the first water quality measurement value acquisition unit, and after the second time point 5. The aeration amount control device according to claim 4, wherein the inflow water quality value of is estimated. The inflow water quality estimation unit
A rain effect determination unit that determines whether the water quality measurement value acquired by the first water quality measurement value acquisition unit is affected by rainfall,
estimating an inflow water quality value after the first point in time based on the water quality change pattern selected from the water quality change pattern acquisition unit and the degree of rainfall effect when the rainfall effect determination unit determines that there is an effect; Item 6. The aeration amount control device according to Item 4 or 5. a second water quality measurement value acquisition unit, wherein the second water quality measurement value acquisition unit acquires the water quality measurement value of the treated water treated in the reaction tank;
The target aeration amount calculation unit, based on the inflow water quality value estimated by the inflow water quality estimation unit and the water quality measurement value of the treated water acquired by the second water quality measurement value acquisition unit, from the blower to the reaction tank The aeration amount control device according to any one of claims 4 to 6, which calculates a target aeration amount to be supplied to. An aeration amount control method used in a water treatment system that performs water treatment by biological oxidation by aerating a reaction tank from a blower,
a water quality change pattern acquiring step of acquiring a plurality of water quality change patterns according to the conditions at the time of acquiring the water quality information acquired in advance from the time series change of the water quality information of the water to be treated flowing into the reaction tank;
a water quality measurement value acquiring step of acquiring a water quality measurement value at a first point in time of the water to be treated flowing into the reaction tank;
Selecting one water quality change pattern that matches the conditions for obtaining the water quality measurement value at the first time point from the plurality of water quality change patterns obtained in the water quality change pattern obtaining step, and selecting the one water quality change pattern and the water quality an inflow water quality estimation step of estimating an inflow water quality value after the first time point based on the water quality measurement value at the first time point acquired in the measurement value acquisition step ;
a target aeration amount calculation step of calculating a target aeration amount to be supplied from the blower to the reaction tank after the first time point based on the influent water quality value estimated in the inflow water quality estimation step. .

Images